Polyurethane-based conducting polymer blends

Polyurethane-based conducting polymer blends

Synthetic Metals 107 Ž1999. 65–73 www.elsevier.comrlocatersynmet Polyurethane-based conducting polymer blends I. Effect of chain extender K.S. Ho a ...

644KB Sizes 0 Downloads 103 Views

Synthetic Metals 107 Ž1999. 65–73 www.elsevier.comrlocatersynmet

Polyurethane-based conducting polymer blends I. Effect of chain extender K.S. Ho a


, K.H. Hsieh b, S.K. Huang b, T.H. Hsieh


Department of Chemical Engineering, National Kaohsiung Institute of Technology, 415 Chien Kung Road, Kaohsiung 807, Taiwan b Department of Chemical Engineering, National Taiwan UniÕersity, Taiwan Received 28 May 1999; accepted 12 July 1999

Abstract Polyurethane ŽPU. with Ž m-phenylene 4-diaminosulfonic acid: PDSA. as chain extender can deeply influence the properties of its blending with n-dodecyl benzene sulfonic acid ŽDBSA. doped polyaniline ŽPADB. The sulfonic chain extender ŽPDSA. provides an additional probability of creating H-bonding with PADB molecules which can be characterized by IR-spectra. In the presence of intermolecular H-bonding, the conductivity and tensile strength of the PADBrPU blends can be changed, especially when the conducting PADB forms a continuous phase in the matrix at high PADB composition. Thermal analysis also reveals variation of glass transition points of the blends at various composition due to the interaction and different degree of miscibility. Scanning electronic microscopy ŽSEM. pictures of the blends demonstrate a more deformed type of morphology when PDBA is used as a chain extender in the preparation of PU. q 1999 Elsevier Science S.A. All rights reserved. Keywords: Polyurethane; Chain extender; IR-spectra

1. Introduction Due to lots of potential application in electronic devices, polyaniline has been widely investigated in the recent years w1,2x. The conducting form of polyaniline can be easily prepared by doping the emeraldine base ŽEB., which has a half oxidized backbone as illustrated below, with protonic acid resulting in a complex acid salt.

However, most of the conducting doped polyaniline ŽPADB. is not soluble in regular organic solvents, which make them no difference from other insoluble conducting materials like metallic, or carbon black powders. Polythiophene which is another kind of conducting polymers has )

Corresponding author

the similar problem was made soluble in organic solvents by covalent-bonded with a long alkyl side chains to each monomer w3x. With the alkylated side chain, the polyŽalkyl thiophene. can even be processed like regular thermoplastic polymers w4,5x. Combining the doping and alkylation idea, n-dodecyl benzene sulfonic acid ŽDBSA. which is also a strong protonic acid with a long alkyl chain, was adopted to dope with EB in the solid-form w6x or in solvent w7x. And the resulting DBSA PADB can be soluble in regular organic solvents, such as, CHCl 3 , toluene, . . . , etc. The improved solubility makes possible the studies on the conducting polymer blends based on the PADB either by mechanical w8–18x or solution mixing w7x. Kuramoto and Tomita w19x and Cai et al. w20x polymerized the aniline in the presence of DBSA or its salt in chloroform and obtain a homogeneous transparent and green–black suspension solution, which can be a polymer solution for polyblend studies. Owning many H-donator and H-acceptor in the backbones or in the complexed form w21,22x, PADB can create various types of H-bonding with other polar polymers such as polyurethane ŽPU.; polycarbonate, . . . , etc. Some of them are capable of building robust H-bonds between different kinds of polymeric molecules. Our studies are focused on preparing a rubbery-like conducting

0379-6779r99r$ - see front matter q 1999 Elsevier Science S.A. All rights reserved. PII: S 0 3 7 9 - 6 7 7 9 Ž 9 9 . 0 0 1 4 4 - 7

K.S. Ho et al.r Synthetic Metals 107 (1999) 65–73


polymer blends based on thermoplastic PU and polyaniline. Expecting the rubbery conducting polyblend can be a material of antenna applied in cellular phone, an anti-static paper roller in a printer, and a fragile, preventive wrapping bag for packing electronic products. The final purpose is to find out the effects of miscibility on mechanical, thermal, and electrical properties of the conducting polyblends.

PADB chloroform solution at various compositions. The obtained polyblend solution was stirred at least 24 h, followed by casting in a petri dish and dried in hood then dried again in a vacuum oven for at least 24 h. 2.4. Characterization A Bio-Rad FTIR ŽFourier-Transform Infrared Spectroscopy:FT-40. was used to characterize PADB and undoped polyaniline ŽPani.. The scanning ranged from 4000 cmy1 to 400 cmy1 with 16 times of scannings.

2. Experimental 2.1. Preparation of PU

2.5. Testing methods PU-prepolymer was prepared with 2:1 mole ratio of purified MDI Ž4–4X diphenyl methane diisocyanate. to polyol Žpolytetramethylene oxide:PTMO. at 708C under nitrogen gas. The obtained PU-prepolymer was mixed with equal moles of chain extender Ž1.4 Butanediol:1.4 BD or m-phenylene 4-diaminosulfonic acid: PDSA. at 708C until the isocyanate Ž –NCO groups. disappear in the IR-spectrum, followed by drying in vacuum oven for 3 days. The obtained PUs are designated in Table 1. 2.2. Preparation of DBSA PADB One liter aniline and ammonia persulfate ŽAPS. separately mixed with 1 M HCl aqueous solution in two plastic bottle were kept in the refrigerator overnight, then the APS solution was added in drop-wise to the aniline HCl solution and kept at 08–58C for 24 h with continuous mechanical stirring. The obtained polyaniline was isolated by filtration and dedoped in 0.1 M ammonium water by stirring for 12 h followed by filtration. The wet cake was washed by ethanol and dried in vacuum oven for 3 days and ground into powder by mortar. PADB was prepared by mixing the dry Pani powders ŽEB. with DBSA with a weight ratio of 1r2 in chloroform w6x. The mixed solution was filtered and the non-soluble cake was dried and weighted to find out the real concentration of the soluble PADB in chloroform solution. 2.3. Blending A PU solution in mixed solvents of chloroform and DMF Ž N, N dimethyl formamide. Ž1:2. was blended with

2.5.1. Tensile strength The tensile strength was measured by a Tensilon ŽMode: TCF-RC, Yashima Works, Japan. followed ASTM D-412 with a crosshead speed of 10 mmrmin and at least five specimens were used for the test. 2.5.2. Differential scanning calorimetric (DSC) DSC thermograms of various samples were obtained from heating the samples from y1008C to 2008C under N2 purging at 208Crmin by a Du-Pont DSC. 2.5.3. Dynamic mechanical analysis (DMA) The tan d , and EX Žof samples were obtained from a Du-Pont DMA at 38Crmin. The sample was 12 = 2 mm and the damping node and frequency are extension and 1 Hz, respectively. 2.5.4. Measurements of conductiÕity (s ) and normalized s Resistance Ž R . was measured by four-probe method and conductivity was obtained from the formula s s LrŽ RA. where: L is thickness and A is cross-section area; and sn is normalized conductivity which was calculated from sn s srf ; where f means weight percent of PADB in PU. 2.5.5. Scanning electronic microscopy (SEM) Morphological studies were performed by using SEM. Microphotographs were taken of the surface made by fracturing the specimen in liquid nitrogen and then casting it with gold ŽAu. powder. A HITACHI S-800 SEM was used for morphological observation.

3. Result and discussion

Table 1 Designation of PU PU designation

Polyol employed

Polyol molecular weight

Chain extender employed



650 2000 650 2000


3.1. FTIR-spectra The undoped polyaniline ŽEB. which was obtained from neutralization of PADB was characterized by FTIR and shown in Fig. 1. The strong peak around 3400–3500 cmy1 comes from the imino group of Pani and the stretching mode of

K.S. Ho et al.r Synthetic Metals 107 (1999) 65–73

Fig. 1. Infrared spectra of polyaniline before and after doping by dodecylbenzene sulfonic acid.

benzoid and quinoid reveal strong absorptions at 1587 and 1480 cmy1 , respectively. The rest absorption peaks are assigned in Table 2. After being doped by dry DBSA, the IR-spectrum demonstrates a different one, which has the hydrocarbon

Table 2 Assignments of the PADB spectrum


Fig. 2. Infrared spectra of blends with different PANDBrPUŽPTMO 650B. ratios: Ža. 0r100, Žb. 10r90, Žc. 30r70, and Žd. 50r50.

stretching around 2800–3000 cmy1 and the sulfonic stretching absorption at 1185 and 1040 cmy1 as illustrated in Fig. 1. The benzoid and quinoid absorption peaks of DBSA PADB shift to lower wavenumbers after doping due to the created positive charge around the backbones. Other typical absorption are listed in Table 2. The soluble part of PADB was blended with PU with a co-solvent to check for the affect of chain extender on PADBrPU, FTIR-spectra of PADBrPU blends were performed and illustrated in Figs. 2 and 3. Based on the chemical structure of PU, PADB, and chain extender, the most possible interaction is the H-bonding as displayed in Scheme 1a and c. The carbonyl of the urethane group which owns the typical absorption around 1700 cmy1 was used to monitor the H-bonding between same or different types of molecules. The H-bonded carbonyl groups inside PU-molecules can be identified from the splitting speaks of 1734 cmy1 and 1705 cmy1 ŽFigs. 2 and 3.. The 1734 cmy1 absorption represented the free carbonyl and it shifts to 1705 cmy1 due to the intermolecular or intramolecular H-bonding of PU-molecules. The incorporated PADB molecules which also has groups being able to induce

Fig. 3. Infrared spectra of blends with different PANDBrPUŽPTMO 650P. ratios: Ža. 0r100, Žb. 10r90, and Žc. 30r70.


K.S. Ho et al.r Synthetic Metals 107 (1999) 65–73

Scheme 1.

H-bonding can alter the ratio of free or H-bonded carbonyl groups of PU. For PU, which is based on the 1.4 BD chain extender, the H-bonded carbonyl group re-split into more peaks when only 10% of PADB was introduced ŽFig. 2b.. We can imagine the PADB molecules destroy the H-bond-

ing inside PU molecules and created H-bonding between PU and PADB molecules. Basically, for PUŽPTMO 2000B. the only possible type of H-bonding between PU and PADB molecules is the H-acceptor carbonyl group of PU and H-donator imino group of PADB, which is shown in

K.S. Ho et al.r Synthetic Metals 107 (1999) 65–73 Table 3 Comparisons of various properties of PADBrPUŽPTMO.. Blends with 1.4 BD as a chain extender Blends

PADB s sn Tensile Tg Žwt.%. ŽSrcm. ŽSrcm. strength Ž8C. ŽMPa.


– 0.57 8.01 16.0 23.5 74.9 – 0.31 6.15 12.2 19.4 74.9

– 5.7 27.0 32.0 33.5 74.9 – 3.1 20.5 24.4 27.7 74.9

15.00 12.90 8.72 7.58 7.12 – 6.45 4.48 3.11 2.34 2.77 –

y7 y6 y4 – – – y62 y59 y57 y52 – –

Scheme 1a. The sulfo-group of PADB which is another source of H-acceptor is shielded by the long alkyl chain of doped DBSA and cannot be free to form H-bonding with H-donating group. The H-bonding interaction comes more strongly when more PADB was introduced and the resplitting of H-bonded carbonyl group at 1705 cmy1 of PU was enhanced. Even the number of free carbonyl group at 1734 cmy1 was also influenced by the presence of 30% PADB in the blends as shown in Fig. 2c and d. Actually, the free carbonyl groups of PU can be divided into two types. One is called amide ŽI. w23x which forms the totally free urethane group and does not have any H-bonding with other molecules demonstrating a absorption at 1734 cmy1 . The other is the H-bonding between the imine and oxygroups from two urethanes groups w24,25x, which is called amide ŽII.. For PUŽPTMO 2000B., the amide II absorption peak decreases finally when more PADB molecules are introduced as shown in Fig. 2a–d. The strong polar attraction of PADB destroys the bonded between them. When PDSA is used as a chain extender, However, this interaction is somehow enhanced. One possible explanation is due to the lack of carbonyl groups which had been used up during the formation of H-bonds with PADB. And more oxy-groups of urethane are left to create bonding with urethane group. The splitting of the H-bonded carbonyl at 1705 cmy1 when PADB molecules are introduced can be attributed to the various types of imines group of PADB. They include the charged and non-changed imines at least. For PUŽPTMO 2000P., the interaction is oven more complicated due to the presence of urea carbonyl group resulting in different peaks around 1700 cmy1 . One of them comes from the H-bonded carbonyl group of the urea. Because there are two kinds of secondary amine groups of urea, they have a large possibility to form H-bond with other groups which also results in more complicated IR-spectra.


The various types of carbonyl absorption around 1700 cmy1 for 10% and 30% PADB in PADBrPUŽPTMO 2000P. as illustrated in Fig. 3b and c are due to the presence of urea group shifting these peaks to lower wavenumbers. When PDSA was used to replace 1.4 BD as a chain extender, each PU-molecule owns one more sulfonic Hacceptor in addition to carbonyl group. It results in the increasing possibility of creating more H-bonds between PU and PADB as displayed in Scheme 1c. That is why the re-splitting of H-bonded carbonyl was more enhanced for PUŽPTMO 2000P. when only 30% PADB was introduced compared with that of PUŽPTMO 2000B. as illustrated in Figs. 2c and 3c. The additional interaction comes from the introduction of sulfuric group at the chain extender can also change the morphology, thermal, mechanical,and even conductivity of PADBrPU blends. If we specify the spectrum in Fig. 2c, we can find the H-bonded carbonyls of PUŽPTMO 2000B. around 1705 cmy1 also disappear when 30% PADB is added. But for PUŽPTMO 2000P. with the same composition Ž30%., this absorption peak is still significant as illustrated in Fig. 3c. The introduction of PDSA did enhanced the H-bonded between different types of molecules ŽPADB and PU.. 3.2. ConductiÕity and tensile strength The conductivity and tensile strength of various composition of PADBrPU are separately listed in Tables 3 and 4 and the diagram of s and tensile strength vs. composition are illustrated in Figs. 4 and 5, respectively. Theoretically, the s-values will be the same after normalization Ž s divided by composition of PADB. if all the PADB molecules in the blends are connected to each other and all contribute to the transport of charge Žconductivity. as displayed in Scheme 2c. The assumption is reasonable because the conductivity of pure PU Ž10–10 srcm. is negligible compared to that of pure PADB Ž74.9 srcm.. Table 4 Comparisons of various properties of PADBrPUŽPTMO.. Blends with PDSA as a chain extender Blends

PADB s sn Tensile Tg Žwt.%. ŽSrcm. ŽSrcm. strength Ž8C. ŽMPa.


– 0.27 5.17 11.60 16.20 74.90 – 0.71 6.92 14.9 23.70 74.90

– 2.7 17.2 23.2 23.0 74.9 – 7.1 23.1 29.8 33.9 74.9

6.13 4.84 3.91 4.52 4.93 – 1.89 2.11 4.14 5.13 5.55 –

– – – – – – y63 y61 y90 y57 – –


K.S. Ho et al.r Synthetic Metals 107 (1999) 65–73

Fig. 4. Conductivity and tensile strength of PANDBrPUŽPTMO 650B..

Therefore, we can use the ratio of normalized s of blends and that of pure PADB to demonstrate the connectivity of PADB molecules in the PADBrPU blends. According to Table 3, when PU is prepared with the 1.4 BD as the chain extender, the connectivity is higher for shorter PTMO diols. On the contrary, the conductivity increases with the molecular weight of PTMO diols when PDSA is the chain extender in the preparation of PU as listed in Table 4. Interestingly, the connectivity Žconductivity. does not always contribute to the increase of tensile strength of PADBrPU blends as displayed in Scheme 2a. Fig. 4 illustrates the increase of connectivity Žconductivity. results in the decrease of tensile strength of blends when more PADB molecules are introduced. Particularly, when conductivity Žconnectivity. comes to a sharp increase due to the threshold effect of PADB Žpercolation. as displayed in Scheme 2a and b, the tensile strength drops more deeply. It indicates that the percolated PADB domain results in an unfavorable effect on strength. And this effect might come from the weaker interaction between PADB and PUŽPTMO 2000B.. However, for PUŽPTMO 2000P., the situation is totally different. According to Fig. 5, the tensile strength increase with conductivity and percentage of PADB. Due to the stronger interaction between PADB

Scheme 2.

Fig. 5. Conductivity and tensile strength of PANDBrPUŽPTMO 650P..

and PUŽPTMO 2000P., the tensile strength was enhanced when connectivity of PADBrPUŽPTMO 2000P. increases. Both counter ion and H-bonding interaction as shown in Scheme 1b and c contribute to the interaction between the bi-continuous PADBrPUŽPTMO 2000P. structure and contribute to the increase of tensile strength. Note that the sharp increase of tensile strength comes at higher PADB composition compared to the percolation Žthreshold. of composition as illustrated in Fig. 5. It indicates that the contribution of connected PADB domain can instantly respond to the increase of the conductivity. Slightly higher composition above percolation is not enough to significantly influence the tensile strength. The other possible reason why the normalized conductivity of

K.S. Ho et al.r Synthetic Metals 107 (1999) 65–73


Fig. 6. DSC thermograms of PANDBrPUŽPTMO 2000B. blends for various PANDBrPU ratio: Ža. 0r100, Žb. 10r90, Žc. 30r70, Žd. 50r50, and Že. 100r0.

Fig. 8. DMA diagrams of PADB125rPUŽPTMO 2000B. Ža. 0r100, Žb. 10r90, Žc. 30r70, Žd. 50r50, and Že. 70r30.

PADBrPUŽPTMO 2000P. is not be equal to that of pure PADB is because part of the PADB molecules are attracted by the PU molecules and become a miscible phase due to the partially miscible as shown in Scheme 2d. It depends on the degree of miscibility of PADB with PU. In other words, if the individual contribution of PADB-rich and PU-rich domain to conductivity can be measured, the degree of compatibility may be obtained merely from the measurement of the conductivity and morphology.

The thermogram of pure PUŽPTMO 2000B. as illustrated in Fig. 5a reveals a glassy transition temperature around y628C and a melting peak around 258C which gradually disappear when more PADB is introduced. It is understood that the long PTM chain can re-crystallize inside PU molecule during quenching when 1.4 BD is used as the chain extender. And the recrystallization is interfered in the presence of PADB ŽFig. 6d.. This recrystallization is not so significant when PDSA was the chain extender as illustrated in Fig. 7a–d. The DSC thermogram of PADBrPUŽPTMO 2000B. with equal weight ŽFig. 6d. demonstrates an endothermic peak of H-bonding around 1208C and a higher de-doping temperature than that of PADBrPUŽPTMO 2000P. system. However, this high de-doping temperature does not happen to PADBrPUŽPTMO 2000P. system as illustrated in Fig. 7. It seems the high hindrance effect of PDSA as a chain extender cannot protect the DBSA from leaving the backbones at high temperature. And the more flexible backbones of PU with 1.4 BD as a chain extender can accommodate the DBSA molecules and keep it ionic bonded to polyaniline backbone at high temperature. The DSC thermograms of PADBrPUŽPTMO 2000P. as illus-

3.3. Thermal analysis The thermal analysis by DSC and DMA are performed to characterize the miscibility of PADBrPU blends. The glassy transition temperatures ŽTg . of each types of blend were measured by DSC and illustrated in Figs. 6 and 7, respectively. The glass transition points are separately listed in Tables 3 and 4. The DSC thermogram of pure PADB as illustrated in Fig. 6e demonstrates a strong endothermic peak around 1808C due to the de-doping of DBSA, which detaches from the polyaniline backbones, resulting in the sudden decrease of conductivity of PADB.

Fig. 7. DSC thermograms of PANDBrPUŽPTMO 2000P. blends for various PANDBrPU ratio: Ža. 0r100, Žb. 10r90, Žc. 30r70, Žd. 50r50, and Že. 100r0.

Fig. 9. DMA diagrams of PADB125rPUŽPTMO 2000P. Ža. 0r100, Žb. 10r90, Žc. 30r70, Žd. 50r50, and Že. 70r30.


K.S. Ho et al.r Synthetic Metals 107 (1999) 65–73

trated in Fig. 7a–e demonstrate the individual characteristic peaks of pure PU, PADB, and blends. Phase separation might enhance at higher temperature for PADBrPUŽPTMO 2000P. system. The glass transition points of the individual blend are listed in Tables 3 and 4, we found that glass transition temperatures ŽTg . of the blends increase with the introduction of PADB. However, due to the unavailable Tg of pure PADB, we cannot analyze the miscibility merely from the change of Tg of blends. DMA was applied to provide more thermal properties of PADBrPU. The tan d vs. temperature for various compositions are illustrated in Figs. 8 and 9. The peak temperature of PUŽPTMO 2000B. system as illustrated in Fig. 8 increases more slowly when compared to that of PUŽPTMO 2000P. system in Fig. 9. This is another evidence showing the better miscibility between PUŽPTMO 2000P. and PADB because this system now owns more types of interaction as shown in Scheme 1b and c. 3.4. Morphological studies (SEM) Fig. 10a shows the SEM-pictures of pure PADB which has a particle-like structure w8,26x. When PUŽPTMO 2000B. is introduced, the particles still exist as shown in Fig. 10b. However, the particles disappear or are deformed when PUŽPTMO 2000P. is introduced as shown in Fig. 10c. It is because PUŽPTMO 2000P. has a better miscibility with PADB compared to PUŽPTMO 2000B.. 4. Conclusion

Fig. 10. SEM-pictures of pure PADB, PADBrPUŽPTMO 2000B., and PADBrPUŽPTMO 2000P..

The introduction of PDSA as a chain extender in the preparation of PTMO type PU can influence the properties of PADBrPU blends by the additional H-bonding and counter ionic interaction between sulfonic group of PDSA and PADB. And these interaction can be analyzed by the FTIR-spectra of carbonyl absorption. The tensile strength of the blends can also increase with the incorporation of PADB which also contribute to the increase of conductivity. Connectivity of PADB rich domain plus the strong interaction between PDSA-chain extended PU and PADB, contribute to the increasing conductivity and tensile strength of PADBrPUŽPTMO 2000P.. The difference between normalized conductivity of blends and pure PADB can be related to the connectivity and miscibility. It is possible to monitor the miscibility of conducting polymer blends merely from the conductivity and morphology obtained from microscopy if the individual conductivity of PADB- and PU-rich domain can be justified. The thermal analysis of blends also demonstrate a better miscibility for PUŽPTMO 2000P. with PADB. SEM-pictures demonstrate the deformation or disappearance of PADB-rich domain in the PADBrPUŽPTMO 2000P. system, but not for PADBrPUŽPTMO 2000B. system.

K.S. Ho et al.r Synthetic Metals 107 (1999) 65–73

References w1x A.G. MacDiarmid, A.J. Epstein, Faraday Discuss. Chem. Soc. 88 Ž1989. 317. w2x M. Kaneko, H. Nakamura, J. Chem. Soc., Chem. Commun., 1985, 1441 w3x R.L. Elsembauer, G.G. Miller, Y.P. Khanna, E. McCarthy, R.H. Banghman, Electrochem. Soc., Ext. Abstr. 85 Ž1985. 1. w4x K.S. Ho, J. Bartus, K. Levon, J. Mao, W.Y. Zheng, J. Laakso, T. Taka, Synth. Met. 55–57 Ž1993. 384. w5x S.A. Chen, J.M. Ni, Macromolecules 25 Ž1992. 6081. w6x K. Levon, K.S. Ho, W.Y. Zheng, J. Laakso, T. Karna, T. Taka, J.E. Osterholm, Polymer 36 Ž1995. 2733. w7x D.C. Liao, K.H. Hsieh, Y.C. Chern, K.S. Ho, Synth. Met. 87 Ž1997. 61. w8x M. Zilbermen, G.I. Titelman, A. Siegmann, Y. Haba, M. Narkis, D. Alperstein, J. Appl. Polym. Sci. 56 Ž1997. 243. w9x Y. Cao, P. Smith, A.J. Heeger, Synth. Met. 48 Ž1992. 91. w10x O.T. Ikkala, J. Laakso, K. Vakiparta, E. Virtamen, H. Ruohonen, H. Jarvinen, T. Taka, P. Passiniemia, J.E. Osterholm, Synth Met. 69 Ž1995. 97. w11x O.T. Ikkala, T.M. Lindholm, H. Ruohonen, M. Selantaus, K. Vakiparta, Synth. Met. 69 Ž1995. 135.


w12x A.J. Heeger, Synth. Met. 55 Ž1993. 3471. w13x Y. Cao, G.M. Treacy, P. Smith, A.J. Heeger, Synth. Met. 57 Ž1993. 3526. w14x Y. Cao, A.J. Heeger, Synth. Met. 52 Ž1992. 193. w15x Y. Cao, P. Smith, A.J. Heeger, Synth. Met. 57 Ž1993. 3514. w16x A. Andreatta, P. Smith, Synth. Met. 55 Ž1993. 1017. w17x P. Passiniemi, K. Vakiparta, Synth. Met. 69 Ž1995. 237. w18x Y. Cao, P. Smith, C. Yang, Synth. Met. 69 Ž1995. 191. w19x N. Kuramoto, A. Tomita, Synth Met. 88 Ž1997. 147. w20x L.T. Cai, S.B. Yao, S.M. Zhou, Synth. Met. 88 Ž1997. 209. w21x T. Vikki, L.O. Pietilac, H. Osterholm, L. Ahjopalo, A. Takala, A. Toivo, K. Levon, P. Passiniemi, O. Ikkala, Macromolecules 29 Ž1996. 2945. w22x O.T. Ikkala, L.O. Pietilac, P. Passiniemi, T. Vikki, H. Osterholm, L. Ahjopaloc, J.E. Osterholm, Synth. Met. 84 Ž1997. 55. w23x N. Luo, D.N. Way, S.K. Ying, Macromolecules 30 Ž1997. 4405. w24x F.C. Wang, M. Feve, T.M. Lam, J.P. Pascault, J. Polym. Sci., Part B: Polym. Phys. 32 Ž1994. 1305. w25x F.C. Wang, M. Feve, T.M. Lam, J.P. Pascault, J. Polym. Sci., Part B: Polym. Phys. 32 Ž1994. 1315. w26x A.J. Motheo, J.R. Santos Jr., E.C. Venacio, L.H.C. Mattoso, Polymer 39 Ž1998. 6977.